Tinnitus, a mostly whistling, buzzing, or hissing phantom sound perceived in the absence of a corresponding external stimulus, is perceived by about 15% of the adult population, and about 75% of all patients with hearing loss. It can cause severe suffering, and to date, no reliable cure has been found. The causes of tinnitus are still poorly understood. The fact that it persists after section of the auditory nerve indicates that its origin lies within the central auditory system. Based on previous research on tinnitus and somatosensory phantom perception in humans and animals, our hypotheses about tinnitus generation are as follows. Damage to the auditory periphery (e.g. as a consequence of loud-noise exposure or aging), which does not even have to manifest itself in a measurable hearing loss, causes some central auditory neurons to lose their input. It is known from research on somatosensory phantom sensations that brain areas that have lost their input due to peripheral damages start responding to stimuli that are normally processed in adjacent areas. We assume that, just like the area of somatosensory cortex that previously processed sensory input from a now amputated hand, starts responding to touches on the face (which are normally processed in an adjacent area of somatosensory cortex), the areas of auditory cortex affected by hearing loss start responding to hearing-loss edge frequencies. This assumption is corroborated by the fact that psychophysical experiments on patients with hearing loss have shown that sensitivity for edge frequencies increases. The resulting imbalance in neuronal activity is mistakenly interpreted as a sound signal - the tinnitus. As not all patients with hearing loss perceive tinnitus, we assume that activity in extra-auditory structures can regulate neuronal auditory activity and prevent tinnitus perception. A likely candidate for such a structure is the paralimbic nucleus accumbens. This part of the ventral striatum has excitatory connections to the reticular thalamic nucleus (RTN), which in turn can inhibit the medial geniculate nucleus (MGN), the thalamic relay between the inferior colliculus and the auditory cortex assumed to be involved with the direction of attention. In tinnitus patients, NAc gray-matter volume is significantly reduced compared to healthy controls. It seems plausible to assume that the volume- reduced NAc of tinnitus patients cannot exert the inhibition necessary to block the excessive activation that ultimately gives rise to the tinnitus percept. It speaks in favor of the assumption of a connection between tinnitus and the NAc that both tinnitus and activity in the subcallosal area (including the NAc) are modulated by stress, arousal, and sleep deprivation. The aims of the proposed research are 1) to provide direct evidence for reorganization of tonotopic maps in central auditory structures by means of high-resolution functional MRI, both in tinnitus patients and in patients matched for hearing loss who do not perceive tinnitus, and 2) to investigate the role of the NAc in tinnitus by a. comparing NAc volume in tinnitus patients and patients matched for hearing loss who do not perceive tinnitus, using high-resolution structural MRI and voxel-based morphometry, and b. comparing, by means of high-resolution functional MRI, activation in central auditory structures and the NAc of patients with intermittent tinnitus in periods during which they do perceive tinnitus and periods during which they do not. Our predictions are that we will find reorganization of tonotopic maps in all patients with hearing loss, but reduced NAc volume and activity only in patients who do (at the time of measurement) perceive tinnitus. These results would guide tinnitus research into a new direction and open up a new point of intervention by emphasizing the modulatory role of extra-auditory structures responsible for the direction of attention.